A “Survival of the Fittest” Mechanism for Weeding Out Potentially Lymphomagenic B-Cells during Germinal Center B-Cell Differentiation.

During normal T-cell dependent immune responses, activated B-cells differentiate into germinal center (GC) centroblasts, which tolerate simultaneous genomic recombination and rapid clonal expansion in order to produce high affinity antibodies. Upregulation of the BCL6 transcriptional repressor is required for centroblasts to acquire this phenotype, since it can repress DNA damage sensing and checkpoint genes. Genetic lesions that cause constitutive expression of BCL6 are common oncogenic events in human diffuse large B-cell lymphomas (DLBCL) and presumably contribute to malignant transformation by sustaining the centroblast phase and facilitating accumulation of genetic errors. We hypothesized that centroblasts must have evolved a mechanism to rapidly terminate the genomic instability phenotype in order to limit the likelihood of malignant transformation. A recent report (Allen et. al. PMID: 17185562) showed that centroblasts and GC T-cells make direct physical contact for ∼30 minutes during affinity maturation. We wondered whether CD40 signaling from T-cells could disrupt the function of BCL6 within this timeframe. Accordingly, we found in ChIP assays that CD40 signaling in B-cells can disrupt the ability of BCL6 to recruit the SMRT and N-CoR corepressors within minutes, at which time RNA polymerase II moves from the promoter to the exons of BCL6 target genes, histones became acetylated, and mRNA levels increase. We showed that signaling from CD40 to N-CoR was dependent on NFkB, while signaling to SMRT appeared to be related to MAP kinase phosphorylation cascades. Two BCL6 targets regulated in this manner are ATR and p53. Accordingly, although CD40 can promote survival of intact GC B-cells, CD40 induced cell death of centroblasts in the presence of higher levels of DNA damage led to cell death, rather than survival. Washout of CD40 after 60 minutes to emulate transient T-cell contact permitted BCL6 target gene mRNA levels to return to their repressed levels, demonstrating that this is a reversible process, which could allow centroblasts that pass quality control to either continue proliferation or undergo terminal differentiation. In order to identify direct targets of BCL6 subject to this regulatory mechanism, we performed ChIP-on-chip of BCL6 in primary human tonsilar centroblasts on a 24,000 promoter array. 915 genes were identified with a cut-off of 99.9th percentile (FDR<0.026), while 1880 genes were identified with a cutoff at the 99th percentile (FDR<0.12). These natural BCL6 targets are functionally related to DNA damage response, transcriptional repression, protein translation, NFkB signaling and others, re-expression of which may play a critical role in facilitating quality control of centroblasts during affinity maturation. Taken together, these data suggest that transient CD40 signaling in the GC might allow T-cells to “weed out” heavily damaged centroblasts while at the same time promoting survival of only the “fittest” B-cells, which could undergo differentiation or additional rounds of proliferation. Others have shown that sustained CD40 signaling can downregulate BCL6 at longer timepoints, possibly reflecting longer exposure of B-cells to this signaling pathway that might occur towards the end of their cycle through the germinal center. Therefore, CD40 can inhibit BCL6 through two different mechanisms, each with potentially different functions during B-cell maturation. Loss of either mechanism is likely to contribute to lymphomagenesis.